Exploring the validity of the chalder fatigue scale in chronic fatigue syndrome - 1998 Morriss et al

Another weird thing is that if there is a feature a particular person never gets - say sleepiness - then they cannot score 0 for that, only ever 1. The more I look at this it seems to have been created by someone with no understanding of simple arithmetic.

 

That makes so much sense. Why do they not ask us to say how much fatigue we have? Of what relevance is how it compares with any particular point in the past?

 

Maybe they've not caught up with the fact that our numerical system does these days benefit from having a representation for zero. It's been around getting on for a couple of thousand years or so, so they should have managed to bring themselves up to date by now.

 

I think there may have been a confusion when the score was invented. For arthritis we have scores of change and scores of current absolute state. The relationship between the ways they used is quite complex. I suspect with this one they had not thought out which they were wanting.

 

Exactly. Before specifying what needs doing you have to specify what the requirement is.

 

Why does it get used so much when it is obviously crap?

 

It couldn't be that it was specifically designed to provide evidence of the efficacy of CBT on fatigue...

 

Yes, it says that at the top of the questionnaire, but the headers for each of the tick boxes says “usual”... why does it not just say “when last well” in those headers? ?

How easy is it for patients to keep making the mental leap to “when well”, when the headers all compare to “usual”?

 

Just to chip in regarding the wording, this is what I was asked to fill in earlier in the year for the Bristol CFS clinic. Highlighted at the start to compare against when you were well, if you've been ill for a while. Still refers to 'usual'.

 

completely agree @Unable

 

It seems that to fill in the less than usual line you have to be even weller than when well.

 

I have to admit that I have not in the past focused on the CFscore, assuming it to be just another blunt questionnaire instrument that may show some sign of change but of doubtful merit. I am beginning to think there is not a lot of doubt about the merit. It looks to fall somewhere to the left of zero.

I had no idea how minimalist the scored question list is. When does 'how yer doing' become a questionnaire?

 

As an aside, it seems remarkable to me that Morriss et al also used some objective measures to validate two of the constructed "four factors of fatigue", and if I understood correctly, they claim the objective measures correlated with the subjective assessments?

Factor 1 (cognitive difficulties)
"Higher scores on factor 1 were significantly associated with subjective everyday minor cognitive difficulties, subjectively impaired concentration on the CIS-R, and the difference between cued and free recall of word-pairs on the paired associate learning task(*)."

Factor 3 (strength and endurance)
Higher scores on factor 3 were significantly inversely associated with left and right hand-grip strength(**), peak heart rate and oxygen consumption on exercise at peak functional work capacity, and peak functional work capacity at their subjective maximum exertion(***).

But also: "There were also significant associations between higher scores on factor 3 and the presence of current DSM-III-R major depression or anxiety disorders."

(I didn't have the cognitive endurance to find out whether they specified the claimed assoaciations)

(*) "Free recall is usually considered to require more effort than cued recall so subjects whose free recall performance was relatively worse than their cued recall performance might be considered to show impaired effortful cognition."

(**) "grip-strength for each hand using a dynamometer"

(***) "heart rate, oxygen consumption in a minute (using a PK Morgan Exercise Test System), workload at peak functional work capacity on a Bosch ERG-551 electronically braked cycle ergometer, calculated as the amount of oxygen (in milliliters) consumed in the final minute of exercise per kilogram of body weight“

 

Objective measures will always show some correlation. Everything is correlated to everything to some extent - positively or negatively. So slow walking is correlated to hypothyroidism and to a broken leg but it is not a very good measure of hypothyroidism or broken legs (or depression).

 

Actually specifying "when you were last well" is a bit of an improvement, but most people have, say, more drowsiness at 65 than at 14 so it is still a useless scale for people like me.

If they have to use a comparison rather than an absolute number scale, why not to how much you would like or even a softer "healthy person" or even "normal".

 

I have thought this through a bit more. The confusion stems from a complete mishandling of the score in terms of basic number theory at the Peano level (i.e. common sense). If you want to compare X with Y then either X or Y being more than or less than Z is not what you want to know. And then giving a negative difference from Z the value of 0 is bonkers.

We are puzzled by why this score has not been debunked. That brings me back to the tendency to forget that the great majority of the medical academic community are blissfully unaware that anything like this exists. The psychological medicine community should be aware though. Maybe psychiatrists are not too good at arithmetic.

 

Probably too objective for them. (The BPS ones I mean).

 

@Lucibee and I talked about particular problems with the Chalder Fatigue scale on a previous thread. We looked at NHS data from Collin & Crawley's 2017 paper and PACE data.

#123 through 128 of this thread https://www.s4me.info/threads/a-general-thread-on-the-pace-trial.807/page-7#post-61308

 

In this trial of exercise in lupus co-authored by Peter White, they used three fatigue scales. https://academic.oup.com/rheumatology/article/42/9/1050/1772059

The Chalder Fatigue scores were the only ones to change significantly. Scores on the other two fatigue scales - the Fatigue Severity Score and Visual Analogue Scale - did not change significantly between groups.

The results were interpreted as follows:

It also seems possible that the Chalder Fatigue scores suggested a change in fatigue that was not really a change at all.

The evidence base for GET and CBT in our illness relies heavily on the Chalder Fatigue Questionnaire, because the change in SF-36 Physical Function scores is often unimpressive/non-existent. I think a good paper outlining its flaws, or maybe better, a study demonstrating them, could be helpful.